Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

gpu: quantization refactoring prerequisites #2512

Merged
merged 5 commits into from
Jan 26, 2025

Conversation

dzarukin
Copy link
Contributor

@dzarukin dzarukin commented Jan 24, 2025

This PR updates gemm_with_post_ops implementation with a new way of loading scales as they are going to be affected after refactor. The main issue is that data_type::undef becomes a valid data type for scales, meaning, they were not specified, which would trigger a bunch of issues in the existing implementation.

This PR guards from that situation.

Thanks to @rjoursler for a patch to deal with undefined data types.

@dzarukin dzarukin requested a review from a team as a code owner January 24, 2025 22:45
@github-actions github-actions bot added the platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel label Jan 24, 2025
@dzarukin
Copy link
Contributor Author

make test
disable test_device_cpu

src/gpu/intel/ocl/gemm/gemm_with_post_ops.cl Outdated Show resolved Hide resolved
src/gpu/intel/ocl/ocl_math_utils.h Show resolved Hide resolved
@dzarukin dzarukin force-pushed the dzarukin/refactor_quant_prereq branch from ca82718 to d7c52ef Compare January 24, 2025 23:49
@dzarukin dzarukin merged commit 4ffd446 into main Jan 26, 2025
4 checks passed
@dzarukin dzarukin deleted the dzarukin/refactor_quant_prereq branch January 26, 2025 18:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants